In order to eliminate differences, which are caused by manufacturing errors, between red-component-to-green-component ratios (R/G) of lens modules and between blue-component-to-green-component ratios (B/G) of lens modules, and thus make correct color temperature estimations, a method for calibrating automatic white balance (AWB) is usually introduced to lens module production lines.
In currently available technology, the method usually employed for calibrating AWB includes the following steps: 1) obtaining a sample distribution chart for red component values and blue component values from mass produced (for example, 50 to 200 units) devices under designated calibration light sources (a light source with a high color temperature and a light source with a low color temperature are usually selected); 2) calculating to obtain average values of the R/G values and the B/G values from a sample distribution chart, the average values being used as golden values, and a device that is closest to the golden values being selected to be a golden set; 3) using the golden set to determine a white area for AWB and a color temperature estimation algorithm; 4) calibrating AWB on a device under test, from which the following ratios need to be acquired: the red-component-to-green-component ratio R_set and the blue-component-to-green-component ratio B_set of the device under test under standard light sources with two different color temperatures, and the red-component-to-green-component ratio R_golden and the blue-component-to-green-component ratio B_golden of the golden set under the two different color temperatures; and 5) calculating the gains, i.e., R_gain=R_golden/R_set and B_gain=B_golden/B_set, of the algorithm for calibrating AWB under the two different color temperatures. Devices whose errors are small can essentially be calibrated to adapt to the white area of the golden set by means of the aforementioned method for calibrating AWB.
However, for devices whose errors are small, such as devices whose errors are less than 10%, relatively accurate calibration can be achieved using AWB calibration methods of currently available technology; for devices whose errors are larger, such as devices whose errors are greater than 15%, or even 20%, using the AWB calibration methods of currently available technology may result in calibration gains being too large or too small, making the differences between the white points of the calibrated device and the white points of the golden set too large. This results in poor calibration outcomes, leading to subsequent inaccuracy in color temperature estimations by the device, and thus lowering imaging quality.